In-mold dielectric analysis data (DEA)
Understand what’s happening in-mold
Naturally, the potential for deviation in the characteristics and behavior of any raw material is unavoidable. Moreover, influencing factors are random at best and can range from the likes of fluctuating humidity and ambient temperatures, to the aging of raw materials and storage conditions.
The power of sensXPERT lies in its unique ability to intelligently monitor and adapt to these many complex variables in real time. The starting point here is sensXPERT’s unique take on dielectric analysis (DEA), and this is achieved through the sensXPERT Dielectric Sensor.
The sensor is installed directly into the mold, offering improved sensitivity and the ability to measure a wider range of material properties. In fact, the sensor’s electrode is in direct contact with the material in the mold and transmits an alternating electric field throughout. This causes the material’s molecules to move. As the material cures or solidifies, the flow of this movement decreases while ‘mechanical viscosity’ increases. These variations enable the dielectric sensor to evaluate changes in the material properties, resulting in a rich source of real time data, that is then combined with existing material science information to generate machine learning models.
Ultimately, these machine learning models can calculate quality indicators, such as the degree of cure, the degree of crystallization, a glass transition temperature, flow behavior or other relevant thermal/mechanical properties of the polymer being processed. With this information, sensXPERT radically revolutionizes the way you control your manufacturing process, allowing you to adjust it dynamically and in tune with any potential deviations.
Relevant industry challenges
To understand the true potential of a feature such as in-mold dielectric analysis, it is important to revisit the various industry challenges it addresses. Below are several challenges or pain points that you may recognize:
A lack of transparency
Currently the plastics industry relies heavily on the usual mix of material characterization, modeling, and process simulation originating from a laboratory. That’s not to mention the additional requirement for an experienced workforce to translate that information within a real-world manufacturing scenario. Consequently, quality control of manufactured parts is only possible after they have been produced i.e., what happens inside the mold during the manufacturing process is invisible.
This means longer cycle times with safety contingencies built in, unnecessary scrap production, and inefficient energy consumption. This lack of transparency also makes it difficult to pinpoint the cause of production errors, thereby increasing the time it takes to recommission a production line.
When processed, it is impossible to expect raw material to behave consistently and predictably - even if the material is from the same batch! In fact, there are many unpredictable influencing factors that can affect material behavior during manufacturing e.g., humidity, aging, storage conditions, seasonal changes etc. Therefore, the current practice of relying on material behavior simulations cannot possibly take all these unpredictable influencing factors into account.
The result is usually a case of reality (aka the manufacturing process in real time) not meeting predefined expectations (in line with the simulation). These unforeseen deviations can lead to fluctuations in production quality, an increase in scrap production, and an increase in associated costs.
The manufacturing ‘Black Box’
As mentioned above, a simulation of the material behavior is created ahead of the actual manufacturing process. Thereafter, the quality of the parts produced can only be assessed post-process. So, what happens in the middle is essentially a ‘black box’.
The in-mold process is invisible to the outside world and there is no way of knowing what is happening in real time – there are only assumptions based on the simulations created in a laboratory environment. This lack of information on in-mold material behavior means manufacturers are unable to detect any deviations and are therefore limited in what actions they can take to optimize the manufacturing process. So even if the manufacturing process remains the same, there is always the risk of producing parts that do not meet specific quality control requirements.
Benefits of in-mold dielectric analysis data
Material behavior is now visible
Despite the many unknown and unforeseen factors that could potentially alter material behavior, sensXPERT can now help you detect and monitor those changes, thanks to a rich source of in-mold dielectric analysis data (collected by sensXPERT’s Dielectric Sensor). In fact, these once-invisible in-mold material behavior deviations, can now be visualized in real time on the accompanying sensXPERT Web App.
Dynamic and adaptive process control
Material deviations within the mold are unavoidable, but sensXPERT’s in-mold dielectric analysis data plays an important part in helping you mitigate their effects. Alongside data from other sources, it is all fed into advanced machine learning models that can accurately predict the optimum point of cure or crystallization and other crucial parameters. Armed with these powerful insights, you can unlock an unprecedented level of dynamic and adaptive control over your manufacturing process.
Quality control for each part produced
Crucially, introducing sensXPERT directly into the mold reveals exactly what is happening to the material during the curing or crystallization phase. So even in the event of an unexpected deviation in material behavior, this new level of transparency means you can reactively respond. In other words, you can dynamically adapt your process to ensure the quality of the part produced meets your requirements, regardless of any deviation. Unlike what happens post-process, this new kind of quality control happens directly in the mold, for each part produced, and in real time!
Reduce cycle times
Collecting such a huge volume of real time data during the manufacturing process also has a significant impact on cycle times. It’s what allows sensXPERT’s machine learning models to predict the degree of cure or crystallization for each individual part produced. Unlike the past when manufacturers had to rely on cycle time ‘safety buffers’, this now gives you precise control over the molding process cycle without compromising quality. In fact, sensXPERT could potentially cut current cycle times by up to 30%!
Reduce the production of scrap
Let’s not forget scrap and how sensXPERT’s dielectric analysis data is also one of the key ingredients in reducing current output of defective parts by up to 50%. As with cycle times, by combining real time manufacturing data with information on material science, sensXPERT opens the possibility for you to dynamically control and adapt the process with absolute certainty. This includes your ability to maintain a consistent level of production quality, regardless of any potential deviations in the mold – and, as a result, minimizing scrap production like never before!
Explore real-world use cases for sensXPERT
Improving in-mold transparency in the electrical encapsulation industry
Find out how two companies successfully used sensXPERT on their reaction injection molding processes to boost in-mold transparency and reduce scrap production rates.Find out more